The increasing importance of natural gas as a source of energy poses difficult gas separation design challenges, as the high flow-rate streams recovered from gas fields are at high pressures (typically about 10MPa) and can contain a high proportion of CO₂ (up to 70%). Conventional separation techniques are usually restricted to low CO₂ content or low pressure feeds, and consequently there is a pressing need for an alternative process that is appropriate for such a scenario.

In addition, increasingly stringent regulation of CO₂ emissions has rendered release of CO₂ to the atmosphere more difficult. There is growing pressure for the captured CO₂ to be re-compressed before it may be re-injected and re-used on site, or transported for further use elsewhere. This step is energy intensive, and could potentially change the economic outlook of the separation process, and consequently the entire natural gas production operation.

A process for the separation of gaseous CO₂/CH₄ has been developed, capable of economically treating natural gas feeds at the high pressures and CO₂ concentrations mentioned. The process has been modelled for a system comprising methane, CO₂ and alkane solvent. The separation process model includes absorption and desorption units, as well as heat exchangers and a CO₂ re-compression stage. An advanced equation of state, SAFT-VR, has been employed in this study. This facilitates accurate representation of the complex thermodynamic and fluid-phase equilibrium behaviour of high pressure/ temperature CO₂-hydrocarbon mixtures. The work has been carried out using the modelling-optimisation software gPROMS. The process is optimised by considering an economic objective over a 15-year lifetime. In addition to finding the best equipment sizes and operating conditions, the alkane chain length of the physical solvent is also treated as an optimisation variable, to achieve the most cost effective absorption for differing CO₂ content.

Several case studies are examined. In particular, the impact of including further impurities in the feed is assessed. The effect of recompression costs on the optimal design is also considered. Given the large variability in feedstock over the 15-year lifetime of the plant, process flexibility is quantified.